Refined empirical models for predicting Fusarium head blight epidemics in the United States
A. SPARKS (1), D. Shah (2), E. De Wolf (3), L. Madden (4), P. Paul (5), K. Willyerd (5)
(1) IRRI, Metro Manila, PHILIPPINES; (2) FHB Consulting, Lewiston, NY, U.S.A.; (3) Kansas State University, Manhattan, KS, U.S.A.; (4) The Ohio State University, Columbus, OH, U.S.A.; (5) The Ohio State University, Wooster, OH, U.S.A.
Phytopathology 101:S170

Fusarium head blight (FHB) of wheat causes significant yield and grain quality losses in U.S. wheat. A web-based FHB forecasting system (http://www.wheatscab.psu.edu/) was developed in 2006 to predict outbreaks early enough to apply preventative fungicides. Further enhancements to existing models and the development of new models are underway using an expanded dataset from 1982 to 2009 containing 527 observations over 15 states. Multiple imputation was used to fill missing values in the data matrix. Current model parameters have been updated using a bootstrap sample of the imputed datasets. This analysis suggests that a model currently used for spring wheat may also perform well for winter wheat. Subset selection algorithms identified several other weather-based predictors that may be of potential use for improving prediction models for FHB. A comparison of the models currently in use, updated models, and newly developed models will be presented.